Toward Reduction in False-Positive Thyroid Nodule Biopsies with a Deep Learning-based Risk Stratification System Using US Cine-Clip Images

被引:10
作者
Yamashita, Rikiya [1 ]
Kapoor, Tara [1 ]
Alam, Minhaj Nur [1 ]
Galimzianova, Alfiia [1 ]
Syed, Saad Ali [2 ]
Akdogan, Mete Ugur [1 ]
Alkim, Emel [1 ]
Wentland, Andrew Louis [2 ]
Madhuripan, Nikhil [2 ]
Goff, Daniel [2 ]
Barbee, Victoria [2 ]
Sheybani, Natasha Diba [1 ]
Sagreiya, Hersh [1 ]
Rubin, Daniel L. [1 ,2 ]
Desser, Terry S. [2 ]
机构
[1] Stanford Univ, Sch Med, Dept Biomed Data Sci, 300 Pasteur Dr, Stanford, CA 94305 USA
[2] Stanford Univ, Sch Med, Dept Radiol, 300 Pasteur Dr, Stanford, CA 94305 USA
基金
美国国家卫生研究院; 美国国家科学基金会;
关键词
Neural Networks; US; Abdomen/GI; Head/Neck; Thyroid; Computer Applications-3D; Oncology; Diagnosis; Supervised Learning; Transfer Learning; Convolutional Neural Network (CNN); AMERICAN-COLLEGE; ULTRASOUND; DIAGNOSIS; CANCER; MANAGEMENT; BENIGN;
D O I
10.1148/ryai.210174
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Purpose: To develop a deep learning-based risk stratification system for thyroid nodules using US cine images. Materials and Methods: In this retrospective study, 192 biopsy-confirmed thyroid nodules (175 benign, 17 malignant) in 167 unique patients (mean age, 56 years 6 16 [SD], 137 women) undergoing cine US between April 2017 and May 2018 with American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS)-structured radiology reports were evaluated. A deep learning-based system that exploits the cine images obtained during three-dimensional volumetric thyroid scans and outputs malignancy risk was developed and compared, using fivefold cross-validation, against a two-dimensional (2D) deep learning-based model (Static2DCNN), a radiomics-based model using cine images (Cine-Radiomics), and the ACR TI-RADS level, with histopathologic diagnosis as ground truth. The system was used to revise the ACR TI-RADS recommendation, and its diagnostic performance was compared against the original ACR TI-RADS. Results: The system achieved higher average area under the receiver operating characteristic curve (AUC, 0.88) than Static-2DCNN (0.72, P = .03) and tended toward higher average AUC than Cine-Radiomics (0.78, P =.16) and ACR TI-RADS level (0.80, P =.21). The system downgraded recommendations for 92 benign and two malignant nodules and upgraded none. The revised recommendation achieved higher specificity (139 of 175, 79.4%) than the original ACR TI-RADS (47 of 175, 26.9%; P,.001), with no difference in sensitivity (12 of 17, 71% and 14 of 17, 82%, respectively; P =.63). Conclusion: The risk stratification system using US cine images had higher diagnostic performance than prior models and improved specificity of ACR TI-RADS when used to revise ACR TI-RADS recommendation. Supplemental material is available for this article. (C) RSNA, 2022
引用
收藏
页数:9
相关论文
共 35 条
[1]  
[Anonymous], 1999, EN 1015-3
[2]   3-D ULTRASOUND AND THYROID CANCER DIAGNOSIS: A PROSPECTIVE STUDY [J].
Azizi, Ghobad ;
Faust, Kirk ;
Ogden, Lorna ;
Been, Laura ;
Mayo, Michelle L. ;
Piper, Kele ;
Malchoff, Carl .
ULTRASOUND IN MEDICINE AND BIOLOGY, 2021, 47 (05) :1299-1309
[3]  
Bertasius G, 2021, Arxiv, DOI [arXiv:2102.05095, DOI 10.48550/ARXIV.2102.05095]
[4]   The incidence of thyroid cancer by fine needle aspiration varies by age and gender [J].
Bessey, Laurel J. ;
Lai, Ngan Betty K. ;
Coorough, Nicholas E. ;
Chen, Herbert ;
Sippel, Rebecca S. .
JOURNAL OF SURGICAL RESEARCH, 2013, 184 (02) :761-765
[5]   Management Thyroid Nodules Seen on US Images: Deep Learning May Match Performance of Radiologists [J].
Buda, Mateusz ;
Wildman-Tobriner, Benjamin ;
Hoang, Jenny K. ;
Thayer, David ;
Tessler, Franklin N. ;
Middleton, William D. ;
Mazurowski, Maciej A. .
RADIOLOGY, 2019, 292 (03) :695-701
[6]   A systematic study of the class imbalance problem in convolutional neural networks [J].
Buda, Mateusz ;
Maki, Atsuto ;
Mazurowski, Maciej A. .
NEURAL NETWORKS, 2018, 106 :249-259
[7]   Quantitative Framework for Risk Stratification of Thyroid Nodules With Ultrasound: A Step Toward Automated Triage of Thyroid Cancer [J].
Galimzianova, Alfiia ;
Siebert, Sean M. ;
Kamaya, Aya ;
Rubin, Daniel L. ;
Desser, Terry S. .
AMERICAN JOURNAL OF ROENTGENOLOGY, 2020, 214 (04) :885-892
[8]   Radiomics: Images Are More than Pictures, They Are Data [J].
Gillies, Robert J. ;
Kinahan, Paul E. ;
Hricak, Hedvig .
RADIOLOGY, 2016, 278 (02) :563-577
[9]  
Hahn Lewis D, 2017, Cancers Head Neck, V2, P1, DOI 10.1186/s41199-016-0021-x
[10]   2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer The American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer [J].
Haugen, Bryan R. ;
Alexander, Erik K. ;
Bible, Keith C. ;
Doherty, Gerard M. ;
Mandel, Susan J. ;
Nikiforov, Yuri E. ;
Pacini, Furio ;
Randolph, Gregory W. ;
Sawka, Anna M. ;
Schlumberger, Martin ;
Schuff, Kathryn G. ;
Sherman, Steven I. ;
Sosa, Julie Ann ;
Steward, David L. ;
Tuttle, R. Michael ;
Wartofsky, Leonard .
THYROID, 2016, 26 (01) :1-133